Odds Ratios are useful in summarising the effect size or magnitude of a difference in a 2x 2 table.
For example
Group |
Sober |
Drunk |
|
|
|||
||||||||<style="TEXT-ALIGN: center"> Complied || 106 || 9 || || || ||||||||<style="TEXT-ALIGN: center"> Refused || 59 || 56 || || ||
Odds Ratio = (106/9) / (59/56) = (106x56) / (9x59) = 11.18
The compliers were 11 times more likely to be sober than the refusers.
You can obtain a confidence interval for thso dds ratio [FAQ/oddsr:using some simple calculations] or by clicking on risk under statistics in the crosstabs procedure in SPSS.